BeyondMasks: Video Object Removal Benchmark for Causal and Physical Consistency

BeyondMasks evaluates whether video object removal methods also remove causal and physical aftereffects like shadows and reflections. The dataset repository supports research accepted to ECCV 2026. It was created by authors from multiple institutions and last updated on 2026-06-26.

Use Cases

• Benchmarking video object removal methods based on their ability to handle shadows and reflections.
• Training models to detect and remove physical aftereffects of objects in video sequences.
• Evaluating the causal consistency of video editing tools based on the described evaluation criteria.

Strengths

• Dataset is associated with a peer-reviewed publication accepted to ECCV 2026.
• Focuses on a specific, advanced evaluation criterion for video object removal.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Freshness

Last updated 2026-06-26 20:07:31; freshness should be verified.